Five of the tools on my old list of AI tools for UX/UI have gone already, Tome and Magician among them. So this list of AI tools for designers has a month on it, and I fully expect half of it to look old by Christmas!
For this one we went through 36 tools, checked each of them on the maker’s own site this week and kept 24. That’s 20 proper picks plus 4 bonus ones at the end. Every screenshot was taken this week, so what you see is what you will get when you click the link (give or take a cookie banner). Where a tool has a free plan I say so, and where it doesn’t I say that too, because nothing is more annoying than signing up for something and hitting a card form on the second screen.
For screens and prototypes
1. Figma agent
If you already design in Figma, start here. Seriously. The Figma agent sits inside Figma Design and builds layouts using your team’s real components, not some random library it made up. You can ask it for a few directions for a screen, compare them and keep the one you like. It came out of beta last week, according to Figma’s release notes, and it runs on AI credits, so the free Starter plan gets a daily allowance and that’s it.

2. Figma Make
Figma Make is the other Figma one, and it’s a different animal. You type what you want, or drop in a frame from your file, and it turns it into a working prototype you can click through and publish to a link. It’s great for testing an idea with real people before anyone writes a line of code. This week Figma added a new Google model to it on every plan, which tells you how fast this stuff moves.

3. Google Stitch
Stitch is Google’s answer, from Google Labs. Describe an app by text or voice and it streams screen ideas onto a big canvas while you steer it. Google rebuilt it around an agent in May, and you can push the result into Google AI Studio for a link you can share. Fun fact for the nerds: it used to be Galileo AI, which was on my old list too.

4. Claude Design
Anthropic launched Claude Design in April. You describe what you need (a screen, a prototype, a one-pager, a few slides) and Claude builds a first version, then you keep refining it by chatting. It can export to Canva or PowerPoint, or hand the lot over to Claude Code to build for real. It’s included in every Claude plan, the free one too, though a big job eats into your usage limit fast. The launch post shows what it does.

5. Framer
Framer now calls itself a design agent, and the AI does the boring bits of a website, the pages and the CMS, then you design over the top. The newest trick is Skills, where you write down your brand rules once and the agent keeps to them (have a look at the updates page). The free plan gives you 500 AI credits and a Framer web address, which is plenty to see if you like it.

6. Relume
Got a client brief and a blank page? Relume turns the brief into a sitemap first and wireframes after, which is exactly how I’d want a junior designer to work anyway. It also has loads of ready-made components for Figma and Webflow. There is a free forever plan that includes the site builder.

7. UX Pilot
UX Pilot makes wireframes and finished-looking screens from a prompt or a rough sketch. The bit I’d look at is Autoflow, which builds a whole user flow rather than one lonely screen, and it syncs both ways with Figma. You can start for free.

8. Visily
Visily is aimed at people who aren’t designers, like the founder or the person in marketing who has an idea for an app. Drop in a screenshot or a sketch and it gives you an editable screen, then it can wire screens into a clickable prototype on its own. The free Starter plan gives you 300 AI credits a month. Nice for getting an idea out of someone’s head before they book your time.

For when you want it actually built
9. Lovable
Lovable builds a working web app from a chat, with a proper database behind it and a link you can send to people. For designers it’s one of the quickest ways to go from “here’s my prototype” to “here’s a thing you can actually sign up to”. The free plan gives you 5 build credits a day (up to 30 a month), and their changelog moves so fast it’s hard to keep up.

10. v0
v0 is from Vercel, the people behind Next.js. It writes React code from a prompt, and the Design Mode lets you nudge things visually instead of arguing with the chat. When you’re happy, it syncs to GitHub and goes live on Vercel in a click. Free gets you 7 messages a day (pricing here). If this end of the list is the bit that excites you, our Vibe Coding course is basically built for you.

For pictures, posters and brand work
11. Midjourney
A couple of years ago I said Midjourney fascinated me, and it still does. It’s still the one for mood and style, the images just have a look the others struggle to match. It now has an Edit model too, so you can change part of a picture by describing it rather than starting again (it’s on their updates blog). There’s no free plan, so you’ll be paying from day one. Have a scroll through Explore to see what people are making with it.

12. ChatGPT Images
The image tool inside ChatGPT got a big update in September. The good news for designers is the text: it can do a poster or a menu with words that are actually spelled right, which wasn’t true not so long ago. You can also leave a comment on one part of the image and it changes just that. It works on the free plan, with tighter limits.

13. Gemini (Nano Banana)
Silly name, very good tool. Nano Banana is the image model inside Gemini, and it’s brilliant at keeping the same mascot or product looking the same across a whole set of edits. The basic version is free in the Gemini app, and the newer Nano Banana 2.1 comes with the paid Google AI plans.

14. Adobe Firefly
Firefly is Adobe’s AI studio for images and video (and now sound too). The interesting change is that it isn’t only Adobe’s own models any more, you can pick ones from Google or OpenAI from inside the same app. There are free daily generations with no subscription, so it costs nothing to try.

15. Photoshop Generative Fill
The one most of you have already used without thinking of it as “an AI tool”. Select an area, type what you want, and Generative Fill puts it there on its own layer, so you can bin it if you don’t like it. It’s perfect for stretching a photo to fit a wide banner. Adobe has made the results a lot sharper lately. It comes with Photoshop, or a free trial. If you want to get properly good at Photoshop, AI bits and all, we teach Photoshop one-to-one.

16. Ideogram
Ideogram has always been the one that gets the lettering right, so it’s good for posters and rough logo ideas. The new Ideogram 4.5 is all about editing: you can keep changing an image turn after turn without it slowly going mushy, which is what used to happen.

17. Recraft
Recraft is made for designers, and the big difference is vectors. You can make a set of icons or illustrations in one style and edit them as real vectors, not flat pictures. One catch on the free plan, and it’s a big one: the pricing page says free images belong to Recraft, sit in a public gallery and can’t be used commercially. Fine for practice, not for client work.

18. Krea
Krea is a browser studio for images and video, and it’s quick. Krea 3 came out the very day I wrote this, an editing model that changes the scene around a product while keeping its label and logo in place (the launch post has examples). The free plan gives you a daily allowance with no card needed, and Krea 3 itself is live on the site now.

For motion and research
19. Runway
Runway is the AI video studio a lot of designers reach for when a still image isn’t enough. Animate an illustration, or edit footage just by describing the change. You get 125 free credits that don’t expire, which is enough for a proper play. Their changelog is worth a look.

20. Dovetail
Not every AI tool is about making pictures. Dovetail takes your interview recordings and notes and pulls the themes out of them, which is the slow, boring part of user research. You can then ask it questions about what your users said. There’s a free plan with one project. If research is your thing, we did a whole post on AI research tools for UX teams.

And 4 bonus ones
Bonus 1: Mobbin
Strictly speaking Mobbin is a library, not an AI tool, but hear me out. It’s a huge collection of screens and flows from real apps (when I wrote this it said 621,500 shipped screens), and now it has an MCP, which lets your AI tools search it and pull in reference screens (that part is on the paid plans). Basically, stop your AI designing from nothing.

Bonus 2: Spline
Spline is for 3D on the web, and the new version has an AI agent that makes 3D objects and textures from a description. Heads up, the free plan doesn’t include any of the AI, that starts on the cheapest paid plan.

Bonus 3: pen.dev
This one used to be called Pencil, and now it’s pen.dev. It’s a design canvas that coding agents like Claude Code or Cursor can draw on directly, and it can export HTML and CSS. Very much for designers who are half developer already. There’s a free plan, though it only gives you 5 days a month of designing with an agent.

Bonus 4: Uizard
For old times’ sake! On that old list I called Uizard my go-to for turning ideas into designs. Miro bought it, it still works and it still has a free plan with 3 AI generations a month, but I couldn’t find a single bit of news from them in months. I’d try the newer ones above first, to be honest.

Which one did I miss? There are new ones every week, so tell me in the comments what you’re using and I’ll check it for next time. And if you’d rather learn this with someone next to you, our AI for UX/UI & Product Design course runs part-time in the evenings, in Bangkok or live online, and the tool list usually includes ChatGPT, Claude, Gemini and Figma’s AI features. If you’re newer to Figma itself, Master Figma With Us is one-to-one.